Despite significant advances made in reducing gay men's involvement in HIV risk-related sexual behaviors and the incidence of new infection in this population, there has been little progress made in reducing young gay men's risk for HIV infection. Prevention efforts have in part failed because significant gaps exist in our understanding of the broad array of individual, familial, social, and community characteristics that both individually and collectively influence young gay and bisexual males to engage in both risk and protective factors. Building on our own research with young gay and bisexual men, as well as other high risk youth, Childrens Hospital Los Angeles is proposing to conduct theory-driven (i.e., the social ecological model) longitudinal research to examine young gay and bisexual males' use of drugs and involvement in HIV risk and protective behaviors in the context of individual experiences (e.g., sexual identity and coming out experience; racial/ethnic identity) and the family (e.g., family composition and environment; familial support/rejection), which in turn will be examined in the context of the youth's interpersonal relationships (e.g., number and nature of intimate relationships; peer ties and integration; peer support), and community contexts (e.g., integration and linkage with community of residence, racial/ethnic community; religious organizations; school; workplace; the gay community; and social service agencies). The proposed study will involve two phases of research. During Phase 1, formative ethnographic research will be conducted to further characterize and contextualize the constructs of interest and the proposed relationship between these constructs and the proposed dependent or outcome variables, and refine and/or adapt existing survey instruments that are appropriate for use with the target population. During Phase 2, longitudinal epidemiologic research will be conducted by following a cohort of 880 ethnically diverse young gay and bisexual males, ages 18 to 22, for 30 months (at baseline and then at 6, 12, 18, 24, and 30 months following baseline); subjects will be recruited using a venue-based probability sampling design, stratefied by race/ethncity and generational status. Longitudinal ethnographic research will also be conducted with a subset of 200 subjects recruited from the larger cohort; data collection will occur at baseline, 12 and 24 months.